Sivananthan S

Experienced Data Annotation Specialist with hands-on expertise in 2D and 3D labeling for autonomous driving and computer vision projects. Skilled in performing LiDAR, cuboid, bounding box, polygon, and polyline annotations with high accuracy and attention to detail. Contributed to large-scale projects such as Galaxy, Vehicle, and Puls, involving object detection, segmentation, and classification for AI model training and validation.

Key Responsibilities:

Performed cuboid, polygon, and polyline annotations for road scenes and traffic environments.

Created LiDAR and bounding box annotations for multiple vehicle types (cars, trucks, buses, trailers, vans, cycles, motorbikes, and kickstart scooters).

Annotated road edges and centerlines using polyline tools for lane and road structure mapping.

Labeled pedestrians (adult and child), static items, and road cones for scene understanding.

Marked ignore areas to refine dataset quality and improve AI model training performance.

Ensured dataset consistency, quality control, and adherence to project annotation guidelines.

Tools & Skills:

Labeling tools: CVAT, Supervisely, Labelbox, Scale AI, or proprietary annotation tools

Annotation types: 2D Bounding Box, 3D Cuboid, Polygon, Polyline, Semantic Segmentation, LiDAR Point Cloud

Domain: Autonomous Vehicles, Computer Vision, AI/ML Datasets

Additional skills: Quality Assurance, Data Validation, Annotation Guideline Review

Projects Worked On:

Galaxy Project: Polygon and polyline annotations for vehicles, roads, and infrastructure objects.

Vehicle Project: LiDAR and bounding box annotations for vehicles and ignore areas.

Puls Project: Cuboid annotations for multiple classes including vehicles, pedestrians, and road elements.

Experienced Data Annotation Specialist skilled in 2D and 3D labeling for autonomous driving datasets. Proficient in LiDAR, cuboid, bounding box, polygon, and polyline annotations with strong attention to accuracy and project quality.

Key Expertise:

3D Cuboid and LiDAR annotations for vehicles, pedestrians, and road objects

Polygon and polyline labeling for road edges, centerlines, and lane structures

Bounding box and ignore area marking for dataset optimization

Familiar with tools like CVAT, Labelbox, and Supervisely